Although you can invert the order of the subprocesses in the application of the model, it is quite strange that you design the frame (2.4) only after having designed the data collection methodology (2.3 that includes also the questionnaire design). You need to know at least who will be your respondents to prepare the questionnaire, and, maybe, you need to know your approximate sample size to decide how to collect your data (also to take into account costs). Thus we propose to reverse the order of 2.3 and 2.4 subprocesses, or, at least, specify in the description of the subprocesses that they are usually performed in parallel.
Change the ordering of sub-process 2.3 and 2.4. Especially objects, populations, and frames need to be considered early.